How big the market has become
The numbers alone justify the attention. Bain projects the global retail media market will reach roughly $140 billion by 2026, according to prior coverage in this series. CPG brands already direct 39 percent of total ad spend through retail media networks, and industry expectations point to that share climbing to 50 percent. That is not a niche reallocation. It is a structural shift in how the industry buys media.
Amazon sits at the centre of that story. Its grocery business crossed $150 billion in sales in 2025, a scale that gives its retail media network commercial weight no pure-play ad platform can replicate. When a shopper's purchase history, delivery address, and search behaviour live inside the same platform that sells the ad, the targeting logic is fundamentally different from anything available in traditional media.
But scale alone does not make spending productive. The central challenge in retail media right now is not access. It is evidence.
Where the measurement gap bites
Incrementality measurement is the discipline of separating sales that a retail media placement actually caused from sales that would have happened anyway. Most brands cannot do that calculation reliably, placement by placement, today. The attribution methods retailers provide most often count impressions that reached existing buyers who were already going to purchase. That is not advertising at work. That is a paid confirmation of a decision already made.
The gap matters more now than it did two years ago because budgets have grown large enough to distort category economics. When 39 percent of a brand's total ad spend flows through retail media, and that spend is measured by retailer-supplied return on ad spend (ROAS) figures that include baseline sales, the apparent performance looks strong while the true incremental return may be much lower. You are, in effect, paying for visibility against shoppers who already chose you.
This is not a theoretical risk. It is the pattern that emerges when brands do post-hoc audits of their retail media portfolios against clean holdout tests. The headline ROAS and the incrementality-adjusted ROAS can differ by a factor of two or more. Until your measurement framework can produce that adjusted number, you do not know what you bought.
On-site, off-site, and in-store: three channels with different problems
Retail media is not one channel. It is at least three, and each has a distinct measurement challenge.
On-site search and sponsored product placements sit closest to purchase intent. A shopper searching for a specific product category is already in a buying frame of mind. The incrementality question here is whether your ad moved them from a competitor's product or just confirmed a choice they had already made. Brands with strong baseline market share in a category are most exposed to the over-attribution problem here.
Off-site display and programmatic, served by retailer data outside the retailer's own properties, is a different proposition. The targeting is better than standard programmatic because it is anchored to actual purchase behaviour. But the measurement loop is longer, and the conversion signal is harder to close cleanly. Brands allocating significant off-site spend without closed-loop measurement against actual basket data are working with weak signals.
In-store digital media, including digital shelf-edge screens, entrance displays, and end-cap digital units, is the fastest-growing and least-measured segment. Physical placement drives genuine impulse and discovery, particularly for new products and new shoppers. But in-store media attribution almost always relies on store-level sales lifts measured across all shoppers in the store, which blends media-exposed and non-exposed buyers into the same result. That blending inflates apparent performance in high-traffic stores and obscures true causal impact.
What retailers are doing with the money
Retailers are not passive participants in this expansion. The largest networks are actively building closed-loop infrastructure and pitching the measurement story as a competitive differentiator. Amazon's advantage, connecting ad exposure directly to purchase, is the benchmark every other network is working toward.
Walmart Connect has invested in its own measurement partnerships and is extending off-site capabilities through its DSP. Kroger Precision Marketing uses Kroger loyalty data to close the loop between ad exposure and in-store purchase. Tesco's retail media offering, built on Clubcard data, gives CPG brands access to panel-level purchase history that most media buys simply cannot match.
The investment signals from French grocery are a useful counterpoint. A French Senate inquiry found that the country's four dominant grocery chains, Leclerc, Carrefour, Intermarché and Coopérative U, working through three buying groups, capture a disproportionate share of value in the food supply chain. That structural power dynamic shapes how brands negotiate retail media terms, not just trading terms. The retailer who controls the shopper data, the shelf placement, and the media inventory sits in a strong position at every point of the commercial relationship.
How brands are activating and where the gaps show
The brands making the most disciplined use of retail media right now share a few practices. They run holdout tests at the campaign level, not just at launch but routinely, to build a library of true incrementality benchmarks by retailer, format, and category. They treat retailer-supplied ROAS as an opening data point, not a conclusion. And they separate new-shopper acquisition from existing-buyer retention as distinct objectives with different acceptable cost structures.
New-product launches represent the clearest case for retail media investment because the incrementality question almost answers itself. A shopper who buys a product that did not exist six months ago is, by definition, an incremental sale. Brands that concentrate retail media spend on new SKUs, new categories, and new geographies, rather than defending existing share in established lines, will generate cleaner proof of return and build a measurement track record that justifies the next budget conversation.
The brands that are struggling are the ones treating retail media as a tax on shelf access rather than an addressable media channel. That framing leads to spray-and-hold strategies where budget goes to on-site search across every SKU to maintain visibility, with no prioritisation and no measurement discipline. At 39 percent of total ad spend, that approach is expensive. At 50 percent, it will be ruinous for brands that cannot show what they bought.
What to watch in the second half of 2026
Three things will shape the retail media conversation through the rest of this year.
First, standardisation. The IAB and MRC have been developing retail media measurement standards for two years. Any movement toward cross-retailer incrementality definitions will change how brands can compare performance across networks, and will remove one of the largest barriers to mature budget allocation.
Second, the Amazon effect on grocery rivals. Amazon's $150 billion grocery business means that any brand selling food in the US is, to some degree, running a retail media strategy against Amazon's network whether it intends to or not. Rivals including Walmart, Kroger, and Target are all investing to close the measurement gap that Amazon currently holds. Watch Walmart Connect in particular, given Walmart's recent commitment to 7,200 price rollbacks, a signal that the retailer is competing hard on basket economics and will use media revenue to subsidise price investment.
Third, in-store digital maturity. The in-store segment is where the next wave of retail media budget is likely to land, because it is the last place that shopper attention has not been priced at scale. The measurement infrastructure for that channel is weakest. Brands that invest now in building in-store measurement frameworks, through controlled store tests and loyalty-linked attribution, will be better positioned when in-store media budgets become material enough to demand accountability.
The $140 billion market is real. The returns are real in aggregate. But aggregate numbers hide enormous variation by placement, by retailer, and by brand situation. Your job right now is not to be present across all of it. It is to know, with confidence, which parts of it are working for you.